Figure 1: Trends of size of global capital markets

Fraud, defined as criminal deception for unjust advantage, has evolved with technology (Bolton and Hand 2002).
Fraud detection usually works along side fraud prevention, where there is a necessity for detection methods when prevention fails.
Classic prevention methods include:
Financial market trading
Illegal trading practices, including insider trading, closing price manipulation, and spoofing, undermine the financial market’s integrity.
There’s a growing body of empirical evidence highlighting various illegal trading strategies.(James, Leung, and Prokhorov 2023)
Spoofing and closing price manipulation are both forms of market manipulation but they differ in their methods and objectives.
Spoofing is an especially prevasive problem in US stock markets where 97% of orders are cancelled before they trade (Khomyn and Putniņš 2021)
JP Morgan paid over $900Million in fines for spoofing activity in the commodities markets during 2008-2016 (Debie et al. 2023)
- This is an actual FINRA manipulation case from 2016(Zhai, Cao, and Ding 2018)
Tradition detection approaches assume that insider trading moves prices because the insider’s private information is reveal to the market through the trading process
These approaches then use principled time series econometrics techniques such as ARMA(1,1)(Park 2010) or structural break analysis of a linear regression capital asset pricing models (Olmo 2011)
The limitation of these approaches is that they are designed to detect long-lived insider (illegal) trading over several months ahead of corporate announcements or unexpected news releases
More sophisticate anomaly detection algorithms such as Nearest Neighbour Dynamic Time Warping (James, Leung, and Prokhorov 2023), Ensemble Gaussian Mixture Model(Emmott et al. 2015), an Isolation Forest and One Class Support Vector Machine.
In a recent research collaboration with Citi Bank’s FX surveillance team we built a principled BERT model, which help to reduce their false positive rate for alerts.
\[ATV(\%)=\frac{\text{No. of rejected Null hypothesis}}{\text{Total number of announcements tested}}\]
Finally, in financial year 2019/20 the FCA introduced the Potentially Anomalous Trading Ratio (PATR).
PATR examines trading around announcements with significant price changes (PPSNAs)1.
It focuses on accounts demonstrating anomalous trading compared to their historical behavior.
Anomalous behavior includes not typically trading in the instrument, trading significantly more in the direction of the announcement, and significant profits from positions established just prior to the announcement.
The ratio is calculated as the value of trading activity considered potentially anomalous during the Pre-Announcement period over the total trading activity in the same period